At the 2006 Austin Game Conference, Rob Pardo from Blizzard Entertainment kicked off the conference with an in-depth look into Blizzard's design philosophy. He illustrated how a few key mantras contributed to the success of World of Warcraft. This is a transcript of his speech.

Rob Pardo is vice president of game design at Blizzard Entertainment. Previously he was the lead designer of World of Warcraft. In 2006, he was named by Time Magazine as one of the 100 most influential people in the world.

Speech Transcript

Thank you all for coming and hopefully I can talk a bit about how we developed World of Warcraft and it will be interesting to all of you.

The point of my talk is to talk about how we develop all Blizzard Games.

We have a core group of philosophies and approaches that we take and what I'm going to do is go through some of those philosophies and how we applied them to WoW.

I'll be talking a lot about some of the different features we put in Wow and also the design philosophies. We have a lot of goofy mantras, like Purity in Purpose, Concentrated Coolness, and obvious ones like Easy to Learn and Difficult to Master. One of the reasons why we have these mantras is, when you have a large studio and you have a lot of designers it is important that every one understands your values. When you're developing that design culture, if you don't have the shared values it is hard for all the designers and developers to understand what you're trying to achieve.

When I first came to Blizzard Allan Adham, founder and President of Blizzard at that time, was also one of the primary designers. He would draw a donut on the white board to show what Blizzard was all about. When you look at the donut the middle of the donut is the core market. Unlike a lot of the companies that talk about just the core market or just the casual market, we feel it is important to make games for both. One of the things that is interesting from the market perspective, as the donut grows, the whole market grows, but the casual market grows much faster than the core market, so it becomes increasingly important to be able to appeal to both of those markets.

One of the chief ways that we do this is through our system requirements. Game development is all about developing cutting edge technology and making better graphics but a lot of time, in that search for the holy grail of realism, you start cutting out that casual market so we've made sure our system requirements are lower so we appeal to a much broader group of players.

The first mantra of Blizzard is definitely "Easy to Learn - Difficult to Master" and going back to that donut, that is really how we achieve that. The fist thing we always do is design the depth first and the accessibility later. I think this actually somewhat unintuitive because a lot of companies are much more concerned about the causal market and they talk about it and talk about it, but we actually do the opposite when we first develop our games. The first thing we do is try to come up with the really cool things. What are the things that are really going to attract players to the game and get them to play the game for 2 or three years? What's our re-playability? Where is the depth coming from? And then we start talking about the accessibility. How are we going to make this approachable and easy to learn, but it all starts with the depth first.

In WoW, some of the things we talked about early on, as far as depth goes, is character classes. When it comes to a class based MMO, I think one of the most important things that you do is develop your combat system. This is the thing that players are going to be doing more than any other thing, so we spent a lot of time early on talking about what combat should feel like and what classes should feel like. And one of our huge driving goals of our characters classes is to try to make them as unique and different as possible. This goes back to when we were developing Starcraft with completely different races, we really want the classes to be the same way.

Dungeons was another big one. This appeals much more to the center of the donut hole. We knew we wanted the dungeons to be a much more hardcore experience. We wanted to have scripted encounters. We wanted to have only groups that could go in there and it's really interesting when you start making content for these different players because we have a lot of players who play solo and really wanted to be able to get into those dungeons. We get suggestions all the time for solo dungeons or duo dungeons. "I want to experience that content." For us, the dungeons are there to serve the core market and if you stay with the game long enough and start developing friend and social networks you can start going into those dungeons. It is something you can strive for and it becomes a bridge for casual players to start becoming more hardcore and part of the core market.

PvP was another big depth decision that we knew we wanted to do. All the previous Blizzard game since Warcraft: Orcs and Humans have been online competitive games, so we knew that a lot of our market was going to be extremely interested in PvP. And we really wanted to make sure the PvP system was fun inherently. Early on we really didn't know how the honor system was going to work, we didn't know if we were going to have titles and achievements but we knew we had to have PvP and we knew that it had to be fun. We also knew it had to be Alliance vs. Horde because that was part of the whole Warcraft brand.

There is a whole group of designers at blizzard that played a whole lot of Ultima Online and Everquest and led uber-guilds and really had all these ideas of what we could do with our raids to make them more interesting and make them have encounters like something you might see in Zelda - scripted encounters. We talked about what worked well in the raid game and didn't work well in the raid game.

After we're done with depth, we start talking about accessibility. Accessibility always start with the user interface. We spend a lot of time talking about the user interface and I think one of the first pitfalls that happens when people develop the user interface is that they try to make everything visible from the interface itself. We try to streamline our interface and we try to present the information that is really important - the HUD, but there are a lot of other things we put into the interface that are difficult to get to, but it is actually part of the art itself. For example getting to our auction house. I see a lot of MMOs or games where the auction house might be a button in the UI but it increases the complexity of the UI itself and the UI becomes less accessible. The more complexity you add to the UI, the more difficult it is for players to learn and explore it. We made the decision with a lot of our UI conventions to make them on NPCs in the world, like the auction house. You go to a city like Stormwind and the auction house is actually accessible completely through and NPC. This layers the UI and makes it easy for players to learn it over time. The only thing you want in your front UIs are the things the players need.

Another thing we talked about early on as a goal for WoW: we wanted the game to be solo-able to 60. We really wanted it to be available to everyone. If you want to start playing the game and play it more like a single player game - you could do that. The idea for us is that as you level up - you're going to see a lot of other content. You're going to see dungeons. You're going to start getting in involved in PvP or battlegrounds. You might start seeing players in really prestigious gear, but none of that happens if you get to level 20 and then stop playing because your character is no longer viable. So we looked at soloing 1 to 60 as our casual game. It still takes hundreds of hours to get to 60 for the casual players, but they're enjoying it the entire time because they're questing, they're getting involved in the story and overall their characters get powerful.

The other thing we spent a lot of time on is the newbie experience. There's a whole lot of games that have a really deep and interesting tutorial but for us we generally shy away from tutorials, even though the notion of giving the player all the information you can so they're armed and ready, tutorials aren't usually very fun. I enjoy games like Prince of Persia or God of War where they start the action right away and start giving you information as you go along to increase your skills and mastery of the game and that is much more the approach we take with our games. When you first start in WoW, we drop you right into a newbie Zone. The thing that's really good about having and open field and one building there is that you're not right in the middle of a huge confusing city There's a lot of other games that we saw where you're newbie experience is basically trying to find your way out of your starting town and we wanted to avoid that.

The other thing that we thought was really important was getting you right into the action. When you start in your newbie zone, there's a couple of NPCs, a building or two and then there's monsters. Within about 5 minutes of starting up your game in WoW you can start fighting monsters. The other thing we thought was really really important that we didn't see in the MMO genre was a game that was completely driven by quests. We want to make sure that you always had a reason for existing, you had a story. A lot of MMOs had problems early on because people didn't know how to get the quests. You would have to go on the web and find out, where are the quests? Its all about Thottbot, we make jokes about Thottbot all the time, but we all use it. The exclamation point became important. This is something we did for the fist time in Diablo 2 and it worked really well to draw player's eyes to what they should be doing, where they should go. So when you first get dropped into the newbie zone, you're right in front of a guard that has an exclamation point above is head. We found that even the most casual players go "Wow, I need to go over there" They click right on it and right away they have a quest. Right away you know what you're supposed to be doing and you're off and running.

On to the next mantra which is: "Killing with a Purpose" which really is the quest philosophy for World of Warcraft. The first thing we talked a lot about, with other MMO's approaches to doing quests, was a lot of the other MMOs you go out and the whole reason for playing the game is seeing that experience bar move. You kill monsters you look at how big a tick you got. Did I get a bubble? Did I not get a bubble? Did this monster give me more than that monster? Did I die a lot to this monster or that monster? That was a big part of the game-play, getting experience. That's really fun game play, but it's not particularly accessible. It's something the more hard core piece of us would be analyzing, figuring out and finding interesting. What we thought would be cooler for most people is if we gave you a reason for killing monster other than watching the experience bar move. Example, the red leather bandana quest. All you need to do is go out and kill bandits and you'll collect enough bandanas to turn in for a reward. A lot of people criticized how many bounty quests or collection quests there are in WoW, but that really came out of the quest design of Killing with a Purpose. Rather than going and finding a spot at your level range that has the most efficient monsters to kill or the most experience, we would give you a reason to kill all the monsters in your level range and give you more quest experience than just killing monsters themselves You're always moving around the world, seeing a new place, seeing a new monster, experiencing different things in your combat.

Another thing that we thought was really important for our quests was clear concise directions. This goes back to, "don't drive players to websites for your game." We try as much as we can to provide all the information to play within the game. We try to tell you not only what it is you need to do but where you need to go and where the quest giver is so that when you go out there and complete the quest you know where to go back to. So these are things we thought were important. Every time we bring in a new quest designer there's the idea that they're going to the 'mystery quest', the really cool quest that you have to figure out where you have to go. "There's some NPC in Stormwind, that give you some piece of information and you need to go into Stormwind and find", but the reality is all the player is going to do is go to Thottbot and find out where the NPC is and walk to him. The people that don't go to Thottbot are your casual players, the people you should be holding the hand for. One of the first things we have to break new quests designers of is the whole concept of mystery quests.

We didn't think it was particularly fun to go into a town to click on every NPC until you find something cool to do. We try to show you the quests. If you go into a town you see all these exclamation points, you can quickly get a menu of options of what you want to do. There is a negative side effect to this. Every time we 'quest-up' a town you'll see what we call the Christmas tree effect. All these little exclamation point Christmas lights all over the town - it's overwhelming. There's balance between putting player on rails, which is one exclamation point, versus ten, which would be too many. We try to strike a balance in between. We try to provide a player with just a few so they have a few options, and as they complete those more will open up. Then we get to the limit of 20. This is one of our most requested quest feature, to raise the cap on the max amount of quests. There are technological limits, the more quests we add the more queries we have to do and that could have an adverse impact on performance. The more meaningful reason is a design reason. We found that the bigger quest log gets, the less you feel that you actually are on a quest to do something. This goes back to the Christmas tree effect. We're actually alpha testing our expansion right now, so our towns are all Christmas tree-ed up. You go into a town and you vacuum up the quests, you have 15 quests in your log and then you don't even read the quest info, you don't pay attention to why you're being sent out to acquire this thing. You go out into the zone and you start killing stuff knowing that chances are with 15 quests, you're probably doing one. We found it really important to put a limit so that players have to make some interesting decisions on which quests they want to accomplish. You put that limit in there then people are now making decisions, there is almost some game play in the quest log itself. "Do I really want to abandon this quest? Am I really going to do it or do I want to do this other quest?"

One of the things we say to our quests designers is that they are the cruise directors of WoW. It is the job of the quest designers to show you the world. We spend a lot of time talking about how we are going to quest-up a zone. We talk about Points of Interest, or POIs. We try to make sure we have quest coverage over all of the POIs. How many bounty quests are there? How many collection quests are there? How many epic quests are there? That's really the job of the quest designer. The quest designer is there to entertain all of you when you're playing the game, and entertainment comes in a lot of different shapes and size. Some people like collection quests, some people don't. Some people like dungeons quests. Some people like to do fed ex quests, or travel quests. The quest designers really are trying to give you that cruise ship agenda and itinerary of all the activities, you might want 80s night or go on an excursion. Same idea. Itinerary, A menu of options so you can pick your own entertainment. You can pick the quests you want.

The next thing I think is incredibly important to any game, especially an MMO game with thousands of hours of game play, is pacing. I look at pacing as the bridge between depth and accessibility. Once you have all those deep interesting game features in your game, you have to figure out how you're going to get people from that newbie experience to that core experience. It becomes all about pacing. For WoW, that really includes the leveling curve. When I hire designers for Blizzard, giving you guys the inside track now, one of my pitfall questions that I ask designers is, "Why do you think Wow is so successful", and one of the hidden answers is the leveling curve. It is somewhat unintuitive to a lot of people in our industry because we want to extend out that content, but once you extend that leveling curve too far it starts working against you. It starts creating what we call 'experience walls'. It might start at 20 where the pacing is falling apart and you're not getting rewarded enough for your time. It might be level 30, 40 and we certainly have some of those walls in WoW, but our walls are shorter and there is less of them which is why we've been so successful. The other thing that is really interesting about getting the level curve is that players re-roll characters and start over. There's a lot of interesting times, especially toward the end of our development where a lot of people on our team were panicking. We were in our alpha and beta tests and some hard core players that would get to level 60 in a week or a week and a half. We had people all across the company that were like, "OMG, what's going to happen. We can't have them doing that. We have to extend the content. We can't have people pay one month and unsubscribe from our game." I would always tell them "No no no, we can't design our game for the guy that is splitting his character with another guy and playing 24 hours a day." You can't balance your level curve for that guy. You have to let him go. If he unsubscribes that's okay, but he's probably not. He's probably going to stay in the game. He's going to start hitting end game content. Or, and this is exactly what happen for us, he'll re-roll a character. He'll have multiple 60s and that's ok. The other thing that's really good about encouraging players to re-roll through the leveling curve is that they'll actually get a new experience. There's all kinds of times that I've played previous MMOs and I might be maximum level. I'll be fighter and I'll say, "You know I wonder what it'll be like to be a mage but I'm not doing that level track again so forget that." Another thing that we did is that we introduced the rest system. The idea for the rest system was really to help even more with the difference between the casual player who might play 4 or 5 hours a week vs. the uber-hardcore guy who plays 20 hours a day. The rest system was meant to bridge the gap between those different play styles. The hardcore will play the game with no rest state so they're having to play a little bit slower vs. the more casual players who take a lot of days off or want to have a weekend binge, they get rewarded for doing that.

The other thing that's really important to pacing is bite-sized chunks. There is a whole host of content in MMOs, and WoW is no exception here, that might require you to play 3 or 4 hours at a time, but one of the things we try to do as best as we can is make sure that even if you're just playing for a half-hour that you're getting something meaningful accomplished in the game. We try to tune our quests so you can accomplish them in chunks. There are all kinds of times that I talk to people who say they log in at lunch time and do a few quests and then go home and play at night. That was a conscience decision that we made.

Another thing that we're doing a lot more in the expansion, that we stumbled on in development so we didn't do a lot of, is winged-dungeons. We split dungeons into separate wings and each of the wings could be done in a half-hour to an hour vs. spending 3 hours to complete. That was a lesson that we learned during development, but since dungeons take so long to make, we weren't able to re-do all our dungeons. You'll see that in our expansion; we're doing a lot more of that. It is really trying to provide that experience so that players can accomplish things in smaller blocks of time, but they can do multiples too. You'll see people who like to do the same wing twice in a night, or all four wings in a night. So you're providing that menu of options based on how a player wants to play, how much time they want to devote to the game. You want to avoid them getting to a point where your game doesn't allow people to play unless they have X amount of time per night. We want to keep those chunks as small as we can. Another example is our battlegrounds. Warsong Gulch was originally called our 'lunch time battleground'. We had a lot of designers that would go play counterstrike or Battlefield 1942 over lunch. We liked that concept, so in our PvP deign we tried to do bite-sized chunks.

The next mantra. Concentrated Coolness. What concentrated coolness means is, rather than make as much variety and many different things as you can, make sure the things you do make are as cool as possible. The place that most illustrates that in WoW are the character classes. Character classes in a lot of RPGs: there will be split classing, advancement schemes, specialized classes, you might have 20 class choices and we consciously avoided that because there are only so many ideas, only so many things you can put in your combat system. We chose to make each one of our classes as cool and different from each other as possible, for a lot of reasons. First and foremost, just so players can recognize the classes and when you played a new class it felt like a new experience. Instead of "Well, this is kind like that other class but with this other side of it." We wanted each class to be viable and different from one another.

Because we concentrated coolness like this, we're able to have unique spells and abilities and mechanics. We don't have to have a lot of spells that are identical across classes but look different. Blue fireball, red fireball, white fireball and that's somehow unique to the class. And even examples like mechanics you might see across class, like we have two pet classes, the hunter and the Warlock, but the way they utilize their pets are entirely different, even the way they train their pets. Again we put all those classes on the white board and listed their central mechanics and tried to avoid sharing those mechanics. That's even something that we recently announced that shamans and paladins are switching sides. For a long time the horde were the only side that had shamans and the alliance were the only side with paladins. The primary reason we un-did that is because we found ourselves starting to merge the paladin and shaman mechanics because we had to keep PvP balanced. That goes against our core philosophy of concentrated coolness so we decided it was less important for each of the sides to have their own class than it was for each of the classes to be as individual as possible. More classes is not always better.

One of the other things you see in a lot of MMOs or RPGs for that matter, or RTSs or any other game, is that once you get enough different combat classes or enough different units in the game - players can only remember so much and understand so much. So if you have 50 character classes and you see one walking down the street you don't even know what that classes does unless you've played it. It matters in a lot of cases. If you want to form a dungeon group, which classes come into your group? Which classes are healers? Which classes are good healers? You lose track of the strengths and weakness and once you get into PvP combat it becomes infinitely worse, because once you see a class on the battleground you have to know instantly what can that class do to me? Even if you did have infinitely cool ideas and you could make your classes completely different from one another, you still shouldn't make too many character classes.

A lot of our initial class idea came out of Warcraft 3 because we'd been developing WC3 for one to two years before we started doing WoW, so a lot of our ideas around the classes came out of the heroes from WC3. Three of our warrior heroes were the Mountain King, the Blademaster and the Tauren Chieftain. In WC3, each of the heroes only have 4 spells. One to one they don't translate well to a deep RPG combat system. Rather than take all the heroes and make them classes in WoW, we concentrated them. We took cool abilities like thunder clap from the Mountain King, critical strike from the Blademaster, and shockwave from the Tauren Chieftain and out them in the warrior, which we thought made the warrior a much cooler and deeper combat class than if we'd taken all the WC3 warrior heroes which were very cool and visual but would have given us too many classes that were too alike one another with not enough unique mechanics. We decided to concentrate that coolness in the warrior rather than spreading it out.

Next I want to talk about trade-offs in game design. I think all game designers are greedy by nature, we want to have the perfect deign. We want to have every cool feature in the game. We want to serve every market. We want casual, core players, moms, dads, dogs, cats, playing together, but the reality is that every design decision comes with a trade off. I always tell my designers that nothing in game design is black and white, it's all shade of grey, compromises and trade-offs. Whenever we can, we generally try not to compromise. Compromise usually results in both sides being unhappy with the result. This goes back to when I was talking about dungeons and the dungeon philosophy in WoW. If we chose to go ahead and make solo dungeons and duo-dungeons, the players who went through time playing in those dungeons are going to feel like their experience was cheapened. The time they put into making a group, delving down into Blackrock Spire and seeing those cool encounters - it's not really special anymore. We wanted to make sure that there were dungeons that we more of a hard core feature so we chose very specifically not to have solo instances, not to have solo dungeons, because doing that would have been a compromise and ultimately both sides would be unhappy.

A few specific trade-offs occurred in WoW craft. I talked about system requirements a little bit, but this a great example. Crysis looks like an awesome game, I'm going to play it but my system can handle it okay. They're making the conscience decision to raise their system requirements so they can have beautiful game with amazing lighting effects and that's going to be great for them, but for us we'd rather have lower system requirements. One of the things we had to do was come up with a stylized art style which is awesome because we have some really amazing artists that like to do that but it also generates a fair amount of negative press for us because our screenshots never compare to something like Crysis. It's always interesting during development because all your graphic programmers want to put in more effects because they want to compete with all the others guys out there and the press wants you to compete, but you have to be prepared that if you go with a stylized look and not as high system requirements - your game's screenshots aren't going to compare. That's okay, because every game we've ever released - 6 months before release they're like "Oh, look at Blizzard's screen shots, look at Starcraft, it's 2D, look at Diablo 2, it's 2D, look at WoW, it doesn't compare", but ultimately we've proven that game play is really what matters and if you have cool stylized art, that's enough. And the other thing that's a real benefit about a stylized art style is that it's very resistant to getting dated. We probably won't have to rev our graphic engine for much longer than a lot of the photorealistic games that look dated much faster.

On the cutting edge side, there are benefits there too. On my examples I'm trying to show realistic tradeoffs, because personally as a game developer I would have a lot of fun working on a more cutting edge graphics game. There are benefits to that. You're really fighting against developer psychology by not doing it. As game developers we want to push the cutting edge and make it better, faster, prettier, cooler and telling developers no you can't do that- that's going to be fight during development. Cutting edge is much easier to market. Everyone wants you to do that. The press wants you to do it. Nvidia wants you to do it. Dell computers wants you to do it. Everyone wants you to push the cutting edge. Everyone wants to get on board to help market your game if you do that, it's a huge benefit.

The next one is a game mechanic issue. World size vs. teleportation. My examples here are WoW vs. Diablo. From the very beginning we wanted our world to be very cool. We're calling it World of Warcraft, we really wanted people to experience it. We made a very early design decision to limit the teleportation in the game. The benefit to this is that it makes the world feel more epic in size. There's also the other factors, players get frustrated with it. They start calling it World of Walk-craft, World of Travel-craft. Every sort of craft I've heard by the way. We did do a couple things to make it easier on folks. We came up with the idea of flight taxis, where you can take a griffon or a wyvern-rider to somewhat quickly move through the world. It still takes time, but it is a great compromise because it maintains the integrity of the world. You're still flying through it. If you take a flight from the Barrens to Ogrimmar you see all this scenery and it feels like a real world and achieves that scale. The other thing that is great about limiting teleportation in your game is that it allows areas of your game to be remote. We have an area called the Ungoro Crater which is this really cool "Land of the Lost" area with dinosaurs wandering around and there is very consciously not a flight path there, so you can't fly directly into Ungoro because we really wanted to have that "Land of the Lost" feel. If one of your classes can teleport instantly into Ungoro you lose that. It's no longer a remote area, no longer a cool area. On the teleportation side you get more social connectivity, which is what MMOs are all about. We know its going to be harder for groups. When you log in and I'm in Ironforge and you're in Stormwind and we want to group up, there's a barrier there. It's going to take 5 to 10 minutes. It's more like real life, "Hey, let's meet at the coffee shop. I have to go do some blacksmithing and you're going to do some leather working and then I need to drop off this quest." If we had instant teleportation like in Diablo it would be much easier for us to be connected. It would be easier for us to make friends and group up for dungeons and that is something we consciously decided to go away from.

Another example is prestige gear vs. customizable gear. I picked this one because this is one players ask for a lot. "Why don't you let us dye our armor? Why don't you let us make cool cloaks with cool embroidery?" I think that stuff is really cool. When I played UO I loved that. I loved that you can make your own character with your own unique look or you can have your guild that has the same colors and you wander around in cool uniforms. Really cool feature, but there's only so much art time that you have and only so many artists. We chose instead to concentrate that coolness on making these really cool prestige armor set that came from very specific places and very specific loot tables. What that meant for us - is it allowed the raiders and the more hardcore players to really get that prestige and that reward. A lot of times for hardcore players, the whole point that they're playing as hard as they do, and I know because I'm a hard core player, you want to show off, you want those cool titles, you want to get that cool equipment, you want to get that cool house on the hill. If the gear can just be dyed to look a different color then you lose that prestige, so we made the conscious effort that the best gear in our game comes mainly from raids. We do make some unique stuff in the PvP system, but that's also a hardcore system. If you do customizable gear, everyone has the ability to be more creative and look different and that huge, it's a great win, but you can't necessarily have both. Which is the point - every decision has costs. If you try to have both you'll end up usually being muddled somewhere in between.

I figured that I probably couldn't get through an entire talk about World of Warcraft and the Blizzard design philosophy without talking about 'polish'. Polish is the thing that I always see whenever our reviews come out. Do a word search and I'll see polish 3, 4, 5 times, so I figured I should talk about it. There is this big assumption with polish, that it's something you do at the end. That the reason Blizzard is successful is because we get 6 to 12 month longer than everybody else and that at the end of that process we spend a lot of time polishing polishing polishing. We do get more time, but that's not where we do all the polish. We do all the polish right from the beginning. It is a constant effort. You really have to have a culture of polish. It's really something that everyone has to be bought into. It is something you constantly have to preach. Every time you task a developer or programmer or artist or designer and they give you an implementation that is at 50% you need to go to them and encourage them, "No you need to get this to 90% so that all we have to do is a little bit of polish at the end." If you're doing that from the very beginning, every task and every bug, you have a much better chance of having a polished product at the end, but if you leave it to the end it's going to be a little more difficult. The other thing that happens a lot when you're trying to develop this culture of polish is - why does it matter that the mouse cursor has one frame of lag. Why does that matter? No one is ever going to notice that. And that is invariably true, but it's not the one feature that people notice, it's the thousands. Again, when you go to the game reviews on Blizzard games, they never point out the singular polished feature, they just say it's polished. It's a blanket statement. You have to make sure that every single one of those tasks is as polished as they can make it at the time they do it.

There are a few phases of polish. First is the design process. It starts in the design process. We're in a new era at Blizzard. When I first started at the company, we were pretty much all home grown and we had very few people with the title 'game designer' and we certainly did not hire game designers from outside the company. That's been changing over the last 4 or 5 years now. It is always interesting to me when I get a designer who has experience and they come in and start doing design at Blizzard. They want to work a hundred miles per hour. If we're doing and RTS game they want to do a unit a week. If they're working on WoW they want to start putting in class mechanic after class mechanic. Every time there are all these flaws that are in those issues. It's like, "Woah, you don't have to work that fast, let's think this through." If you just start jamming ideas into the game without thinking them through, without talking to all the other smart folk, smart programmers, artists and other designers in your team, then you're going to find yourself in a world of hurt once you have a 100 features in the game that have flaws, that don't work well with each other. You really have to talk through it in that design process. When we get into a design meeting we try to consider everything. We talk about if it's going to be fun, we talk about solid mechanics that work in the end game, we talk about PvP, is it going to work in this raid encounter, in this dungeon. We talk about the art and cool factor. We involve our art directors and our creative directors. Is this inherently cool? Contrary to popular belief we do consider production. We bring the producers in and talk about how long it's going to take to do this. We try to design smart whenever we can. A great example is a feature people ask for all the time, mounted combat. Mounted combat is very cool, but it's a lot of work and its something that we don't really see a lot of bang for the buck in our game style so that's something we've put the kybosh on mainly due to production. We bounce things off everyone on the team whenever we can. I go to the programmer next door and pitch to him and he'll usually have some negative feedback and I'll go think about that again and then I'll go pitch it to designer, he'll have some negative feedback. I'll go pitch it to an artist. During that process I'm iterating but I'm iterating before I've ever entered my task.

The whiteboard. This is how we start developing our zones. Chris Metzen is the Vice President of Creative Development, he does all our lore and story. He'll come in and draw a map and we'll go from there. The great thing about doing the maps on the white board is that they're easily erasable. We'll move points of interest around, we'll say that coast doesn't look good. We'll change everything right there and there's no cost to the iteration.

Phase two. Now it's time to task. Now we're going to make something. We've done our design doc. We put the tasks in. The programmers are working on the technology. The artists are working on concepts. The first thing we do is try to make it fun. This is an area where we probably do things a little bit differently.

Elwyn forest is the first starting area we did. Northshire Valley is the 1st through 5th level area for the humans and we spent and inordinate amount of time working on that area. We would work on how the combat is going to work? How many monsters can you take at one time? What's the downtime? Does the combat feel fun? What's the first ability you're going to get? What's the pace of the quests? How much experience can you get when you get out of Northshire? Where should we put the trainers? How do we teach you to get to the trainers? We spent a ton of time. We probably spent more time on Northsire Valley by an order of magnitude than any other area because it was one of our first area. We spent a lot of time making it fun first. After we made it fun then we made it big. We didn't go out and build the entire World of Warcraft until we knew what we were building. If we just start building a huge world and haven't figured out all the little details. What's going to make your game fun? What's going to make your combat system fun? How's your quest system going to work? Then once you do figure all that out, you're going to go back through 20, 30, 40 zones and try to retrofit all those ideas back into the game? You're going to be screwed. So we spent a lot of time on making it fun before making it big. And that one of the things that when we went into our friends and family alpha test, that was a whole year before we released the game, everyone was like, "OMG, it's really fun." Like they're all surprised. "You can only play to tenth level, but it was a fun ten levels." But it was a lot easier once we knew what was fun to go from levels 10 to 20 and 20 to 30 and 30 to 40 from there - at that point is was more of an issue of production than design. The design at that point was more creative design than mechanics design.

Another thing I think is really important is game control and is something I think is taken for granted a lot of the time. For the fighting games - brawler genre like God of War, I know it's huge. Nintendo makes it a big deal too. But, I don't hear a lot of game developers talk about control because it's taken for granted. It is really important. It is something your programmers have to be on top of and need to be talking about. There's an example from back when I was developing WC3 where I could feel some lag in the mouse cursor, it wasn't a lot but it was a little bit. I kept going to our lead programmer and saying I think there's something wrong with the mouse cursor. And he kept telling me that "I checked it. I looked at it. It's all fine. It's all good." Over a period of a few months I went to him a number of times. He said, "Okay, I'm going to program in a utility to try to figure it out." So he programmed in a hardware cursor and a game cursor so you could see both at the same time and lo and behold there were three frames of cursor lag. Not a big deal right? But, it is. That is a big deal. Your frame rate is a big deal because these are things that contribute slightly and visually to that game control and if your game control isn't very responsive players will never tell you that. They're not going to mention that in the forums because it's something that's more subconscious. But they will leave your game over it. You'll never know that that's the reason.

Beware of the grand reveal. Kharazan is going to be shipping in the expansion set for WoW, but it was actually supposed to be a dungeon in the original game and the reason that it's not in the original game is because it fell victim to the grand reveal. What the grand reveal is, is when you send off a subgroup, like group of artists or a group of programmers, or a group of designers and they stop interacting with the rest of the team, they don't bounce ideas off everyone else, they just create something in a vacuum. And then three months later they show it to you. And they're all proud of it and excited about it but it doesn't work. In this case, this dungeon was supposed to be a raid dungeon but all the halls were tight and narrow. How is this going to work when I'm bringing 10 or 20 people throughout the hallways. They're like, "But it looks great! It looks awesome." I'm like, yeah, not so much. Why don't you go back and re-do that for the next three months. So it got kicked to the expansion. It is really important that you teach all your developers to show you stuff mid-stream. They're not going to want to do it. They're going to be like, "Yeeeeah, I don't want you to see that its not going to look so good" You gotta say, "No no, that's okay, I just want to give you some feedback to make your life easier as you work on this so you're going to do less iteration on the tail end."

Have fun with the game. Another thing that our universes are know for is that we don't take them so seriously. We have lots of pop culture references. We have lots of funny sayings that the NPCs will say. That doesn't really come from the top; that comes from a culture of everyone adding in this funny stuff. The thing that's really cool about that is that if all your developers are having fun making the game, the players will have fun playing the game too. They'll enjoy all that.

Then the last phase is the phase that really does happen at the tail end which is all the iteration and all the polish. We'll usually pool developers from all the different teams into a strike team. A artist from team one, a designer from team two, we try to put together a diverse group. We get casual players who aren't familiar with the genre. We'll have a guy from the RTS team that doesn't like MMOs to test the MMO content. Then you'll also have all the hardcore players. You'll give them assignments. Have them play the game, play an aspect of the game and then get them back in the room and get feedback. Then we take that feedback and relay back to the game. You're going to uncover a whole host of things through that process, and I know a whole bunch of people here do that process. Some people do it in the beta test is but you can't control the test itself. You can't give people an assignment. You can't ask people probing questions on why they didn't like something or what they didn't understand and often time when the feedback comes unfiltered, what people will give you is feedback that isn't necessarily true, it's their perception of what they think is true. They'll give you what they think is a solution without really understanding why they had the problem in the first place.

This goes back to that - "no one is going to notice one feature but they will notice a thousand." Don't take those small decisions for granted. Especially in that newbie experience. We had cases early on where people would group up with one other person in Northshire Valley we found they would get into the next area which is Goldshire at 4th level and if you get there at 4th level you have a very difficult time very quickly. We try to make sure that we don't take those small decisions for granted, we really try to ask a lot of questions. We try not to let stuff die on the strike team feedback list.

And then of course there's the beta test. The beta test, for us, is not about finding bugs. It's not really about getting a lot of game feedback. It's much more about stress testing the game from a technological and game-play level. We encourage our beta testers to exploit the hell out of the game. We want to see: is there a zone in the game that people level much too fast? Is there a way to get really phat loot ripping through instances? Those are the sorts of thing we use our beta test for. On the RTS side it always funny. We run a ladder in the beta test and we always get a lot of our more competitive players pissed off when we do a ladder in the beta test. "It's not fair." The point of the ladder in the beta test has nothing to do with being competitive. I want the beta test ladder to be full of exploiters. I want to see the top ten exploiters so I can go look at their games. That's what I want.

The last thing. Don't ship until it's ready. I think every game has this where you ship before its ready, you're really going to cripple the chances for success of that game, but for Massively Multi-player games the stakes are much higher. It has a subscription model and you really want your game to have longevity. People generally don't take a second look at your game. There has been a whole lot of MMO games that have shipped early, admittedly so, by those companies. In a lot of those cases you hear great stories now - how much more fun that game is. But how many of you have gone back and looked at that game? No one actually goes back. If you ship before you're ready you're going to cripple yourself. You're putting at risk the next 5 years of your product. Hopefully, all you publishers in the room will make sure you give all the developers more time.

The last thing I want to say is to thank all of you for allowing me to speak and allowing me to be a part of the massively multiplayer genre and I hope that we turn the genre into something special. The thing I think is really unique about MMO games, is look at all the other genres. FPS, RTS, and of all those other cases, the genre takes a very specific type of game-play. Massively multiplayer games - all that says is a game with lots and lots of players. This genre really has the biggest frontier. We should be pushing it in all kinds of different directions.

Thank you for your time.

‍

Browse more of history's greatest speeches →